An optimized software solution for parallel video decoding in digital TV sets

Abstract

This paper presents an approach in software realization of Picture-in-Picture decoder (PIP decoder). Advantage of introduced PIP decoder is that it can be easily integrated into systems without hardware support for Picture-in-Picture functionality. These systems are usually based on a single low processing power hardware decoder, which is in charge of decoding and displaying current video stream. If it is required to decode simultaneously two different video streams, decoder as such one won't be enough. Cheap and simple solution is to use presented PIP decoder which separately decodes and displays another video stream, thus allowing parallel visualization of two video streams. PIP decoder utilizes modified MPEG-2 decoding algorithm, which decodes and displays only certain frames (I-frames) and their elements (DC coefficients per each block of pixels). DC coefficients of each I-frame are collected in order to create a picture. Described approach to PIP decoder realization follows the inability of the target system to use the conventional MPEG-2 decoding algorithm, due to the restrictions of the system to work in real time.